CELL-BASED STRUCTURES

The original note, as written at the time

We need to work our way from down — up. We have always built structures that come out as giant objects, but currently it’s going to be important to develop cell based technologies. Think of intelligent structural systems that respond to the forces about them. Think of intelligent structural systems that can self-repair. Basically every part of the object is sensitive and responsive to use and damage.

Ongoing development

We need to work our way up from below, rather than always building structures that come out as giant single objects.

The important direction now is cell-based technology: intelligent structural systems that respond to the forces on them, and that can repair themselves. Every part of the object sensitive and responsive to use and damage.

A building today is a monument that pretends not to change. It is designed for a load, built for that load, and then it ages in one direction until somebody inspects it and decides. The information about how it is actually behaving exists only in the cracks.

A cellular structure inverts that. The load path is distributed across many small elements, so damage in one is not a failure but a change of state. The structure knows because the cells are sensors as well as struts, and repair is local because the unit of repair is small.

The advantage is not strength. It is that the structure can be maintained continuously instead of being replaced catastrophically, which is the same shift as repairable goods.

The objections are severe and should be stated. A structure with millions of load-bearing cells has millions of ways to fail in a way no code currently describes, and we have no method of certifying it. And the intelligence must fail safe: a cell that stops responding must behave like a cell, not like a hole.

But the current model has its own unearned confidence: we design for a load, guarantee nothing about the third decade, and call the result safe because a calculation said so.

Buildings should behave like organisms in one respect only — they should be able to tell you they are unwell before they fail.